[07:14:51] halfak, you're awake too! [07:14:58] I've been dealing with the silliest data request :/ [07:15:04] Yup. [07:15:05] ugh [07:15:14] silly fun or silly herpderp? [07:15:22] researchers who want some geolocated read logs with country code - TS - URL [07:15:32] I explain this can be used to link back to countries for pseudonymous editors [07:15:36] Sounds like Reid's work. [07:15:39] "ah, but we don't want the edit logs!" [07:15:57] I'm sorry, but if you don't know our edit logs are public you don't know enough for me to spend my time getting your data. [07:15:58] Can you give them a sub-component of the URL? [07:16:06] Yeah... even so. [07:16:11] like, just namespace? nope. [07:16:12] checked. [07:16:25] +1 also, you're at a hackathon and they can wait for their free data support. [07:16:30] Who is the request from? [07:17:10] some guy in the Computer Communications Research Group at USCSC [07:17:14] I've otherwise never heard of him. [07:17:30] I just CC'd you on an email with a researcher @ OII who may want something similar. However, he'll be at the hackathon. [07:17:44] OII == Oxford Internet Institute [07:17:56] Han-teng? already replied :) [07:17:59] yeah, I know em ;p [07:18:19] the thread is actually an interesting example of exploits in anonymisation strategies, though, so it might be worth me bookmarking for when reid thinks about this more. [07:18:38] "ah, but what if you deliberately made some entries WRONG!" "..then I'd expose FEWER pseudonymous users?" [07:18:48] +1 please do share the thread. [07:18:58] * Ironholds nods [07:19:09] Say, interested in demoing WMFUtils? [07:19:09] anyway, I should try and wander downstairs [07:19:23] at the hackathon? [07:19:31] It's probably too limited to internal interest [07:19:53] for any of the interesting functions you need private data and/or access to stat100(2/3) [07:20:01] maybe ua_parsing you can do just with private data ;p [07:20:01] If you think so. No worries. I just thought it might be a good chance to get some users. [07:20:19] the max user base is you, ellery, reid, leila and dario, alas. [07:20:26] gotcha. [07:20:32] So, demos for us later :) [07:20:43] yep. but I might demo it to dario and see if he can spot holes in it or at least refine my demoing technique, because it's not an easy thing to install. [07:20:53] and he's the least familiar of us with R-as-a-general-computing-tool I think ;p [07:21:39] hmmn. [07:21:45] I wonder if this place does a free breakfast [07:23:11] halfak, fancy wandering to find breakfast with me? [07:41:59] Woops. Got side-tracked. [07:42:13] I'm downstairs in the restaurant now. [07:42:15] Ironholds, ^ [07:43:33] * halfak had a random come ask me about SSL and interacting with his bank over shared wifi because I looked like I knew about technology. [07:43:40] :) [09:37:59] halfak: frobisher 1-2 confirmed? I’m going to tweet from @Wikiresearch [09:39:07] Yup. [09:39:07] I was mistaken earlier. [09:39:09] Was looking at a threater. There is a separate conference room. [09:39:15] *theatre [09:39:17] :) [09:40:48] ok cool [09:40:48] yeah that’s the one that I saw [09:40:52] anything particular you want to mention in the tweet? [09:40:52] halfak ^ [09:40:52] maybe the demo session? [09:41:15] Demoing Quarry and Halfak's research libraries. [09:41:24] can link to quarry [09:41:38] http://quarry.wmflabs.org/ [09:42:06] we can make dedicated tweets about this [09:42:45] Sounds good. We can tweet again with the link to the stream [09:42:55] halfak: so official start is 2.25 right? [09:43:28] 1530 [09:43:32] Rather 1600 [09:43:39] We'll be using that first half hour to set up [09:43:45] 1600 BST == 1500 UTC [09:43:51] ah right [09:43:57] kk [09:44:11] 8AM PDT, 10AM CDT, 11AM EDT [09:44:21] Timezones! [09:47:51] Ironholds, the IPs we were using to identify congress edits -- were they from the congress edits twitter bot? [09:51:29] hey halfak! [09:51:33] yes, they were ;) [10:37:46] ping Ironholds [10:37:53] researching stuffs without you. :P [10:38:01] halfak: I'm adding another machine now [10:38:16] Hokay. Think it will take more than 30 min? [10:38:22] halfak: no, shouldn't. 10 maybe/ [10:38:22] ? [10:41:15] halfak: I need to go find Coren to increase the quotas, so I can add moar machines if needed :) brb [10:41:27] Hokay [10:43:14] halfak, pong :) [10:43:21] sorry, I'm downstairs; the humidity was getting to me [10:43:25] but I can wander up if I'm needed [10:46:22] Hello everyone :) [12:24:42] halfak: heya! where you at? [12:34:29] ashaw: https://wiki.toolserver.org/view/MySQL_queries [12:41:04] halfak: https://wiki.toolserver.org/view/MySQL_queries [13:07:41] DarTar: http://etherpad.wikimedia.org/quarry-announce-email [13:08:15] thx [13:08:39] YuviPanda: got some interesting sample queries to showcase? [13:08:46] DarTar: ask halfak [13:08:48] DarTar: :D [13:09:27] are queries and their results permanently cached or do they die after a while? [13:09:30] is the hackathon at 4pm btw? [13:09:39] DarTar: permanently cached, yeah [13:09:43] ok cool [13:22:09] ashaw: http://www.google.com/search?q=mass+email+mediawiki [13:22:30] yes? no? [13:22:50] * Ironholds is currently sat in his hotel room curled in a ball so an answer to when exactly he's expected to spend time with actual people would be nice. [13:25:10] Ironholds: you okay? Hope you uncurl long enough for us to say hi... [13:25:13] Ironholds: there’s no expectation but uncurl yourself and get out of that room [13:25:24] yeah, but I'd like to [13:25:27] Ironholds: maybe just uncurl tomorrow :) [13:25:31] assuming I'm up for it [13:25:38] knowing when I should make that assessment would be useful ;p [13:25:57] ashaw, eh, probably going to have a similar problem though. heat + humidity = miserable. [13:26:57] Ironholds: nothing quite so cool and unhumid as the summer in SF... [13:27:35] well, guess who was stupid enough to decide to move to boston [13:27:36] * Ironholds waves [13:27:39] I'mma have FUN. [13:27:45] but I'll also have /air conditioning and cold drinks/. [13:41:40] anybody else at the barbican finding that SSL for google searches is automatically diabled?! [14:02:31] ashaw, yes. WTF [14:04:53] halfak: yeah, pretty wild. I use duckduckgo by default, so I hadn't noticed until DarTar sent me a link :P [14:05:37] Yuvipanda: quarry rocks (and yes, I'd like to propose that as a new motto/tag-line for the project) [14:05:43] interestingly it only applies to Google, Wikipedia over HTTPS works just fine [14:06:32] yeah, gmail is fine too [14:06:35] super weird [14:17:59] Anybody Philadelphia people here for the virtual day of the hackathon? This is Matt Flaschen. [14:20:11] yo superm401 [14:20:16] Hey, DarTar [14:20:56] ashaw: there’s someone from the press asking about the most recent research on the gender gap, can I point them to you? ;) [14:30:09] oh sure! :) [14:30:52] I'm not a great interviewee, but at least I'm somewhat qualified to answer that question :) [14:36:45] DarTar, can you come help me run the AV setup? [14:36:52] yup, coming [14:39:24] halfak, can you post to the channel when you have the hangout URL? [14:40:30] Yup. Just sent you an invite. [14:42:18] http://youtu.be/iCXlz5cU1q4 [14:42:22] ^ Stream [14:46:36] halfak, ping me if you need help with the hangout. [14:47:47] Scratch that last link. [14:47:47] live demo now on: http://youtu.be/vhRifTmPfSI [14:47:50] ^^^ [14:47:52] This one instead [14:51:12] halfak, I think it's in record mode already; not sure if that's intended. [14:52:22] halfak, DarTar: the demos are not aired, yet, right? [14:52:42] they will be shortly at: http://youtu.be/vhRifTmPfSI [14:52:49] ow I see. this one works. thanks! [14:53:05] np :) [14:53:29] hi leila! are you not in London? [14:53:40] I'm not, ashaw. :-( [14:53:48] leila: :( indeed [14:53:52] are you? [14:54:01] yes [14:54:12] nice! enjoy it! :-) [14:54:36] for sure. [14:55:37] huh. interesting choice to license academic work by-nd [14:55:50] personally, i feel like i want derivatives of my academic work! [14:58:17] jmo shoutout! [14:58:51] HenriqueCrang: I muted you because of background noise from your mike [14:59:12] DarTar, all the mikes are muted now so I can't hear. [14:59:13] DarTar, thanks. I thought I had already done that [14:59:23] leila: any better? [14:59:29] right now everything is muted [14:59:35] or for some other reason, there is no sound [15:00:07] better now? [15:00:12] can't hear anything. [15:00:17] saw HenriqueCrang for a sec [15:00:19] now we have sound [15:00:20] :-) [15:00:22] cool [15:00:22] ok cool [15:00:44] thanks, guys! [15:01:04] leila, :-) [15:11:47] DarTar: should you invite me or should I just use Aaron's laptop? [15:12:12] let me invite you [15:12:20] DarTar: ok [15:12:29] DarTar: wait, let me log in to gmail :) [15:12:31] HenriqueCrang: this seems similar to what we did on https://pt.wikipedia.org/wiki/Module:Avalia%C3%A7%C3%A3o [15:12:50] YuviPanda: looks like I can’t because I’m not the host [15:12:58] HenriqueCrang: ^ [15:13:10] YuviPanda, send me your gmail [15:13:35] HenriqueCrang: ypandian@wikimedia.org [15:13:56] done [15:14:43] "this" = this "wikiclass.classify" thing halfak is talking about [15:15:16] helderwiki, for sure ! [15:22:13] HenriqueCrang, that Lua module is really cool. [15:22:37] :-) [15:24:24] helderwiki, could set up a presentation on it for the global community ;) [15:30:48] YuviPanda, what would it do if you didn't put the LIMIT in? [15:36:00] superm401, it would go throw the role page table, take some time, and give a really big result [15:36:20] I was wondering if it would add an automatic LIMIT. [15:36:29] I 'll ask him [15:36:32] Doesn't seem like it based on the one that lists number of users registered per day forever. [15:54:41] superm401: sorry about the lack of audio. [15:55:12] YuviPanda, it's okay. I could hear everything the presenters say, a good amount of the audience, but not everything. [15:55:15] Sorry for accidentally muting you. [15:55:23] Quarry is very cool. [15:56:04] YuviPanda, think we lost audio from the presenters just now [15:56:10] just now? [15:56:10] DarTar: ^ [15:56:32] ashaw, yeah, I heard the beginning of Contropedia, but it stopped. [15:56:37] aha [15:56:40] the hangout stopped [15:56:42] they're fixing it now [15:56:48] halfak is, i mean :) [15:56:53] back now? [15:57:14] hopefully it's up again [15:58:13] ashaw, yeah. [15:58:26] superm401: awesome [15:58:28] superm401: :D help for developing Quarry welcome! It's just python, and is on gerrit at analytics/quarry/web [16:01:27] Just a heads up - The link to the workin paper (At http://contropedia.net/demo/contropedia_wikisym2014.pdf ) is giving a 404 [16:06:20] Following up to the redirect discussion on Hangout, there is a redirect table (https://www.mediawiki.org/wiki/Manual:Redirect_table), but it says it's not filled out for older articles. [16:06:34] My concern with pagelinks would be that some redirects can have links after the actual redirect. [16:07:26] E.g. https://en.wikipedia.org/w/index.php?title=ArtificalLanguages&redirect=no [16:08:46] all, if you’re on twitter help retweet YuviPanda’s tool for live SQL querying https://twitter.com/WikiResearch/status/497042030452015104 [16:09:18] superm401: you might be interested in this: http://networkcollectiv.es/wiki-redirects/ [16:09:25] DarTar: :) [16:09:29] i'm checking out the link you sent [16:10:12] ashaw, yeah, I'm sure it's harder if you need historical data. [16:10:28] That looks like a good package for that. [16:11:13] superm401: mako knows more of the details of how we handled the really old stuff - i worked on some of the analysis and the writeup for that piece with him - but my recollection is that stuff does get hard if you want to go back in time beyond a certain point [16:11:24] let me go see if i can figure it out... [16:11:59] the code is all available too [16:12:58] ashaw, cool. I don't need it at the moment, I was just commenting in response to the discussion during the Quarry talk. [16:13:07] * Nettrom RT'ed YuviPanda and applauds ashaw & mako for the cool tool and datasets [16:13:16] Nettrom: \o/ [16:13:36] superm401: oh totally! i was just thinking that if there's some limitation, i hope you'll help us find it and we can document it on the paper's website (and release an updated "preprint" version) [16:13:40] YuviPanda: somehow I wasn't following you on Twitter? Fixed that now, though [16:13:42] thanks, Nettrom! [16:13:43] sorry [16:13:50] also, welcome :) [16:13:58] thanks, hello everyone! [16:13:58] Nettrom: :D [16:15:45] ashaw, it looks like that works by parsing each individual revision text, so it can tell when/which revision each page becomes a redirect/changes target/ceases to be one. [16:15:47] http://projects.mako.cc/source/?p=redirect-tools;a=blob;f=01-extract_redirects.py;h=1690941e1ac9290b25ef8947016ef5ee3f9f0a70;hb=HEAD [16:15:49] Seems sound [16:18:18] superm401: yes, definitely. I was trying to figure out if there was some issue with really old redirects... [16:18:26] it seemed like you were implying something like that? [16:19:00] ashaw, I was relaying the issue with the redirect table mentioned at https://www.mediawiki.org/wiki/Manual:Redirect_table . But it doesn't seem like this code uses that, since it looks at every revision directly. [16:19:24] aha. i see! [16:19:29] that's great. thanks [16:20:23] that's the bit about needing to use the pagelinks table [16:21:24] Just finished up. Thanks guys. We'll be in the group hangout shortly. [16:21:45] halfak, side note, http://projects.mako.cc/source/?p=redirect-tools;a=blob;f=01-extract_redirects.py;h=1690941e1ac9290b25ef8947016ef5ee3f9f0a70;hb=HEAD seems to use a similar library to your dump iterator. [16:23:20] http://youtu.be/iCXlz5cU1q4 [16:23:51] Group hangout! https://plus.google.com/hangouts/_/event/csd7v3nephec9gqq0p863dodgrk?authuser=0&hl=en [16:24:08] Anyone who wants to talk research projects, please connect up. [16:26:37] Hey folks, I'm waiting in the (non-broadcast) group hangout. Join up if you want to talk research ideas. [16:28:14] Completed youtuibe broadcast: http://youtu.be/vhRifTmPfSI [16:28:23] We'll cut it up and put it on commons ... eventually. [16:34:10] * Nettrom joined the group hangout [16:34:25] hey Nettrom :D [16:34:52] o/ [16:35:39] Nettrom: tried out Quarry? :) [16:41:14] YuviPanda: haven't gotten around to it yet, loaded up the page so I can give it a go later when I have a spare moment [16:41:21] Nettrom: \o/ sweet [16:41:31] curious to see what you've made, it's an awesome idea! [16:41:47] Nettrom: :D [19:40:28] YuviPanda: is there a plan to add a "preview" button to quarry? I don't want to create garbages when doing try-and-error with queries.